Let the tools be dumb

When Adobe Flash came out, it sparked many minds. People used it to do everything: websites, cartoons, games. Two decades had passed before people realised its limitations.

Any tool trying to do everything gets disqualified and ridiculed. By design, tools have one job they are best for. Other jobs will always call for a different solution, and eventually, people always make one.

For Adobe Flash 15 years ago, such ‘other job’ was serving the smartphone market with its quickly expanding variety of screen dimensions.

These days, GenAI is flicking through jobs frantically testing its boundaries. It’s fascinating to watch!

At work, when we talk about GenAI, I often hear about ‘taking the magic out of it’. Because it’s not magical, it doesn’t do anything you can think of.

Among others, Seth Godin is reminding us to keep our brains switched on to guide the guesswork of LLMs:

? If a guess is good enough, you’re set. If it’s not, plan accordingly.?

See the ChatGPT is dumber than it looks piece for Seth’s short list of use cases where GenAI brings true value in everyday life.

Everyone is in the experimentation phase right now. Experimentation is necessary but costly and produces a lot of waste.

That said, some of us are still disproportionately optimistic, like Des Traynor of Intercom in their first Off Script video. And that’s how it should be. We need both the grand vision and the reality checks along the way.

I’m excited to see GenAI finding those areas where there is truly no better alternative—be that incredible efficiencies or ways to take bias out of conversations or something we haven’t even tried yet.

In the meantime, plan accordingly!

~ ~ ~

Truly yours, Sergey Soloviov